Peltier-based dehumidifying cabinet

PaulS:
Who recommended that? They should be fired.

I'm going by the specs listed here for model #MCPE-241-14-15. Notice, near the top of the page, where it says "Suggested current = 0.7 X Max".

Perhaps it means something other than what I thought it meant?

Current is pulled, not pushed. The cooler will pull whatever current it needs. Your power supply needs to be able to supply that amount of current.

I see. In that case, I presume 6A is simply the minimum the power supply should be able to handle? What's the "suggested current" bit about, then?

First thing you need is a way to measure the temperature, and humidity, in the cabinet.

Already have those.

The cooler may be PWM controllable, or it may not. If not, you would just need to turn it on and off at some interval, experimentally determined, that prevents the device from freezing.

I've read that power cycling isn't good for peltier coolers, in the long run. I've also read that peltier coolers should not be controlled via PWM.

What is the ideal humidity level / temperature for your guitars?

There are two different and conflicting issues here. The first is that a traditional PWM scheme is a bad idea because it will be inefficient. As you increase the voltage, or current through a Peltier the cooling output will increase reach a peak and then decrease due to the additional power dissipated in the Peltier. In a PWM controll scheme you are emulating a lower supply voltage / current by pulsing a higher one on and off. But in this case the higher your input voltage is the less efficient the device will be, so you are much better off having a lower steady input voltage than a higher pulsed input voltage from an efficiency standpoint. That said, if your Peltier is sufficiently oversized you might be able to just write off the efficiency loss.

The other issue with PWM is thermal cycling. Anything heated and cooled enough times will be mechanically harmed by the effects of differential expansion and contraction of materials. What you want to avoid is a rapid thermal cycling of the Peltier device. If you PWM quickly (faster than 10 Hz) no appreciable temperature change will occur in the device during the cycle and you should be fine.

If you can capacitivly filter your PWM signal so that it is essentially a DC signal you will solve both of these issues at one time, so that is the best solution, but a quick PWM with a maximum voltage that is still in the good efficiency range would probably work.

Perhaps it means something other than what I thought it meant?

I think it means that you need to have a power supply capable of supplying the maximum amperage the device needs, without stressing the power supply. A 6A power supply supplying 6A all the time will get hot, and not last as long as you might wish. A 6A power supply supplying 4.2A all the time will be a lot cooler, and last a lot longer.

I want to build a dehumidification cabinet for my guitars and need advice on what parts to get.

First, buy a hygrometer so you know what you're up against (low? high?).
Maybe just need one or two packs of Zorb-it for your cabinet.

robtillaart:
What is the ideal humidity level / temperature for your guitars?

Ideal humidity level would be somewhere around 50%. Mean relative humidity here is 76.5%.

vasquo:
Maybe just need one or two packs of Zorb-it for your cabinet.

Zorb-its are good for maintaining average humidity levels, but not so much for use as dessicants. Average humidity is fairly high here in Puerto Rico, meaning Zorb-its would only serve to maintain our rather high average humidity.

jroorda:
There are two different and conflicting issues here. The first is that a traditional PWM scheme is a bad idea because it will be inefficient. As you increase the voltage, or current through a Peltier the cooling output will increase reach a peak and then decrease due to the additional power dissipated in the Peltier. In a PWM controll scheme you are emulating a lower supply voltage / current by pulsing a higher one on and off. But in this case the higher your input voltage is the less efficient the device will be, so you are much better off having a lower steady input voltage than a higher pulsed input voltage from an efficiency standpoint. That said, if your Peltier is sufficiently oversized you might be able to just write off the efficiency loss.

The other issue with PWM is thermal cycling. Anything heated and cooled enough times will be mechanically harmed by the effects of differential expansion and contraction of materials. What you want to avoid is a rapid thermal cycling of the Peltier device. If you PWM quickly (faster than 10 Hz) no appreciable temperature change will occur in the device during the cycle and you should be fine.

If you can capacitivly filter your PWM signal so that it is essentially a DC signal you will solve both of these issues at one time, so that is the best solution, but a quick PWM with a maximum voltage that is still in the good efficiency range would probably work.

This is very good info. A PWM signal filtered by a capacitor looks like the best approach for temperature control, but what about humidity control on top of it? It seems to me that switching the PWM signal on and off according to measured humidity levels would lead to the same expansion and contraction issues I am trying to avoid. What, then, would be the best way to accomplish the same effect?

Which are you trying to control - temperature or humidity?

Have you researched humidors? An oversize humidor seems like what you are trying to create.

PaulS:
Which are you trying to control - temperature or humidity?

The goal is to control humidity, but the cooler needs to be at the right temperature for water vapor to condense without freezing over.

Have you researched humidors? An oversize humidor seems like what you are trying to create.

Humidors are meant for humidification rather than dehumidification. I want to remove moisture from the air, not add it.

I'm looking to build something like this, but less expensive and without the humidifier component (which I don't need).

So what you want to do is set the cold side of the peltier to the dewpoint so that you can collect the condensation and remove it from the cabinet.

You're going to attach some sort of aluminum plate/heat sink to each side of the peltier. The hot side will need a fan to turn on whenever the peltier is active. On the cold side you will also need a temperature probe of some type (e.g. LM35 or DS18B20) to measure the temperature of that side. Use the PID library to control the power on/off to the peltier (via relay or mosfet, preferably mosfet) to keep the dewpoint temperature of the cold side stable.

There's no PWM needed here, or putting it another way there's no reason why you'd need to switch the peltier on/off more than a couple times a second to keep a reasonably stable temperature.

Chagrin:
There's no PWM needed here, or putting it another way there's no reason why you'd need to switch the peltier on/off more than a couple times a second to keep a reasonably stable temperature.

What about thermal stresses, as noted by jroorda? His reply is consistent with this FAQ, which states:

"Significant precautions must be employed with PWM, however. First of all, the PWM should be at a high enough frequency to minimize thermal stresses to the TE devices. While we like to keep the frequency in the low killihertz (Hz) range, in many applications these days we must compromise at around 120 Hz for the sake of electromagnetic compatibility. Another important issue is the potential for generating electro-magnetic interference (EMI) in the wiring to the TE device. If you are using PWM, you may need to shield your power wiring or keep it away from any sensitive electrical signals."

That FAQ and all of their technical specifications provide pretty much no detail as to thermal stress considerations, and if it were a big issue to be concerned about I would expect that they would list some specification as to the number of switching cycles it can handle. It also contradicts itself with its discussion of using mechanical relays by suggesting that the relay fails before the peltier junction does? After digging around with Google I'm not denying that PWM would be preferable, but I still can't find any decent numbers as to how much of a concern this is.

Stick with the K.I.S.S. principle. A peltier junction is ~$10, and if it does wind up failing too quickly then you can revisit how you're driving it. Drive it at a low voltage and tweak the voltage a bit until you get a reasonably low switching rate. You don't really need that much cooling; At 70F (room temperature) the dewpoint is only 50F for 50% humidity and that should be a much much lower power requirement than these devices are capable of.

Has a long life, with Mean Time Between Failures (MTBF) exceeding 100,000 hours

I don't see you wearing it out really fast you can lower the voltage or current to maintain temperature. Maybe PWM is not the name to use more like turn it on let it cool a big plate with some mass to it so you end up with real slow cycles. You hit 50 turn it off warms up to 55 turn it back on.

I've decided to go with capacitively filtered PWM to control the cooler's temperature while switching it on and off altogether to control humidity. So...

  • What's a good place to buy a suitable power supply for the 29.8 VDC, 6A cooler model? This should be able to power both the cooler and one or two 12V computer case fans (using, I presume, a voltage divider for the fans).
  • What kind / spec of relay should I get that would allow me to drive the cooler via PWM at 29.8 VDC and assuming a current of 6A max?
  • What kind / spec of capacitor should I get for the purposes of filtering the PWM signal above?
  • What kind of sensor do I need to tell if water has reached a certain level inside a container?

Side note: if both of your fans will always be running you can attach them in series and that would make them use 24 volts which is close to your 29 volts.
someone correct me if I am wrong.

arduinopi:
Side note: if both of your fans will always be running you can attach them in series and that would make them use 24 volts which is close to your 29 volts.
someone correct me if I am wrong.

I'm not sure I want them running 24/7 (thinking about power consumption and wear), but perhaps it doesn't matter?

Answering some of my own questions:

  • Mouser carries a large selection of power supplies, so I'll choose one of those.
  • A relay seems to be the wrong choice for use with PWM, but it looks like I could use this MOSFET I found on Sparkfun. Right?
  • For the water sensor, I suppose I could do something like this.

Still hoping for advice on capacitor specs.

No advice on capacitance values needed for a project like this?

Concerning the water level sensor: it turns out the sensor described in the link I posted is subject to corrosion from electrolysis. I've found some circuits (here and here) that turn DC power into AC power on the electrodes in order to prevent corrosion from electrolysis, so those are probably a better choice.

Anyway, thank you all for your help. I may post an update with instructions if I ever get this thing built.